What happens to a shallow clone of a Delta Lake table if VACUUM is executed on the source table?

When a VACUUM command is executed on a Delta Lake table, it removes the old files from the file system that are no longer required for versioning or are beyond the retention period set for the table. In the case of a shallow clone, this operation affects its underlying data because a shallow clone is essentially a link to the original table's data files at the time the clone was created.

As a result, if the source table undergoes a VACUUM operation which deletes certain versions of the files, those files may be removed and become unavailable to the shallow clone. Since the clone relies on the original table for its data, the removal of these files disrupts the functioning of the clone, making it unable to access the data linked to it. Consequently, the shallow clone will not operate correctly as it can end up referencing non-existent files, leading to potential errors or incomplete data when attempting to query the clone.

This behavior highlights the importance of understanding how shallow clones in Delta Lake are directly tied to their source tables, particularly in relation to data management operations like VACUUM that impact underlying storage.